MCP Server for Cursor: Connect ToolYour Tools

Modern software development has evolved rapidly with the rise of AI-assisted coding. Developers using Cursor rely on local agents to generate code and refactor systems. However, an agent's utility is limited by its sandbox environment. Without external tools, agents predict code patterns instead of interacting with live data or executing real-world functional tasks.

By using the MCP server for Cursor, you can give your local agent direct access to an online suite of utility tools. ToolYour offers a remote Model Context Protocol (MCP) server that connects your IDE directly to a rich catalog of 200+ utilities. Whether you need to run SEO checks, perform file conversions, or manage complex text formatting, this integration allows your agent to work with production-grade backend APIs instead of writing brittle, custom scripts from scratch. Learn more at https://www.toolyour.com/developers/mcp.

What ToolYour MCP adds to Cursor

Integrating ToolYour into Cursor changes how your editor's agent handles complex manual tasks. Normally, if you ask an agent to extract metadata from a page or convert file formats, it must write, compile, and execute scripts locally on your machine. This process is slow, prone to compilation errors, and consumes significant token context.

Using the remote MCP protocol, your agent accesses ToolYour's central directory of over 200+ utilities. The agent identifies the correct tool schema, sends the payload, and receives structured data back in milliseconds. According to the open standard specifications on https://modelcontextprotocol.io, this architecture decouples the LLM from tool execution. By moving the processing load to ToolYour's remote server, your editor stays lightweight, and your workspace remains secure from running untrusted local scripts.

Prerequisites (account and API key)

To use the remote MCP server, you will need a ToolYour developer account. While browser-only tools can be tested for free without signing up on the homepage at https://www.toolyour.com, programmatic access via REST or MCP requires authentication.

To set up your environment:

1. Register your account at https://www.toolyour.com/signup.
2. Navigate to your developer credentials page at https://www.toolyour.com/dashboard/api-keys.
3. Create a new API key. All valid developer keys start with the ty_ prefix.

This unique API key is the only credential you need. It manages authorization across all surfaces and tracks your request limits.

Cursor mcp.json configuration

Cursor reads remote connections from your local configuration settings. You can register the ToolYour remote endpoint, which is located at https://api.toolyour.com/mcp.

To set this up, add the configuration details to your client settings. Use the following structured JSON code block:

{
  "mcpServers": {
    "toolyour": {
      "url": "https://api.toolyour.com/mcp",
      "headers": {
        "X-Api-Key": "ty_YOUR_KEY"
      }
    }
  }
}

Make sure to replace ty_YOUR_KEY with your actual key from the dashboard.

To configure this in the Cursor user interface:

1. Open Cursor and go to Settings > Features > MCP.
2. Click + Add New MCP Server.
3. Name the server toolyour, select the sse transport mechanism, and add the URL https://api.toolyour.com/mcp.
4. Add a header called X-Api-Key containing your ty_ key.

Review detailed protocol documentation on the Cursor manual at https://cursor.com/docs/mcp and https://cursor.com/docs/context/mcp.

First agent workflow with solve_task

Once connected, your Cursor agent interacts with the server through a series of specialized meta-tools. Instead of overwhelming your LLM's context window with hundreds of tool schemas, ToolYour uses a discovery and orchestration system.

The meta-tools exposed on the protocol include:

• solve_task: The primary agentic tool that routes tasks to the correct execution utility.
• discover_tools: Lists available endpoints and schemas.
• list_categories: Groups tools by category (such as SEO or converters).
• get_tool_schema: Details specific input and output structures.
• invoke_tool: Directly executes a utility backend.
• load_skill: Imports specialized instructions.
• run_workflow: Executes chained, multi-step integrations.

When you ask Cursor's Composer or Agent to perform a task—such as parsing metadata from a website—the agent calls solve_task. The ToolYour remote dispatcher evaluates your instruction, selects the appropriate backend tool, executes the web parser, and formats the output. This streamlined execution avoids multiple context exchanges and delivers results directly into your code editor.

What tools are available (API-backed only)

ToolYour maintains a catalog of 200+ tools. However, it is important to distinguish between client-side and server-side utilities:

• Website-Only Tools: These interactive widgets run entirely inside your browser and are designed for manual human usage on the main portal. They are not exposed through programmatic environments.
• API-Backed Tools: The remote MCP server only exposes backend, API-backed utilities. These are optimized for programmatic execution, offering predictable JSON outputs for text formatting, converters, and data parsers.

This design guarantees that your agent will only call stable, server-run APIs. You can view developer specifications and schema patterns at https://www.toolyour.com/developers.

Quota and pricing (shared with REST)

Managing development resources is easy because ToolYour uses a shared consumption model. Your REST API and MCP requests use the exact same API key and monthly quota limits.

• Free Tier: Your account starts with a free quota of 500 requests per month.
• Shared Pool: If you make 300 requests via the Cursor agent and 200 requests via a custom script, your total usage is 500 requests.
• Scaling Options: If your projects require more requests, you can upgrade your plan. Learn more on our pricing plans page at https://www.toolyour.com/pricing.

This unified billing setup ensures that you only pay for the raw requests your workflows consume, regardless of the integration surface.

Troubleshooting connection issues with the MCP server for Cursor

If your connection status icon in Cursor turns red, review these common troubleshooting areas:

• Key Validation: Ensure your key is typed correctly. It must start with the prefix ty_.
• Header Format: The custom authentication header must match X-Api-Key exactly.
• Endpoint URL: The target address must be exactly https://api.toolyour.com/mcp.
• Quota Depletion: If requests fail with billing errors, log into your profile at https://www.toolyour.com/dashboard to check if your 500 free monthly requests have been exhausted.

For technical assistance with the open protocol, read the troubleshooting details at https://docs.anthropic.com/en/docs/agents-and-tools/mcp.

FAQ

Can I use my API key for other MCP clients?
Yes. The remote endpoint supports Cursor, Claude Desktop, Claude Code, and other MCP-compatible clients. You can reuse the exact same ty_ prefixed key across any supported environment.

Do tool discovery queries consume my monthly quota?
No. Discovery meta-tools, such as discover_tools and list_categories, are completely free. This allows your LLM agent to search for active features without deducting from your monthly request limit.

Are browser-based tools free without an account?
Yes. Browser tools on our main website are free to use and do not require signup. However, programmatic features require a valid key.

Where do I monitor my request limits?
You can check your remaining balance, view system usage, and manage active developers' credentials on your account dashboard at https://www.toolyour.com/dashboard.

Is there a local installation requirement for ToolYour MCP?
No. It runs entirely as a remote Server-Sent Events (SSE) server at https://api.toolyour.com/mcp, eliminating the need to install or run heavy Docker setups locally on your device.

Conclusion: Empower Your Workflow with the MCP server for Cursor

Integrating the remote MCP server for Cursor bridges the gap between static code writing and active task automation. By giving your Cursor agent access to ToolYour's rich suite of backend tools, you can run complex conversions, data validations, and SEO audits directly within your workspace.

With a generous free tier of 500 monthly requests, a shared API quota with your REST products, and seamless remote setup, getting started is straightforward. Obtain your credentials from the dashboard, configure your local client, and start optimizing your developer experience. For a complete guide, explore the implementation specifications at https://www.toolyour.com/developers/mcp or view the professional tiers on the pricing page at https://www.toolyour.com/pricing.